As global volatility and unpredictability continue to affect organizations, more finance and risk managers are recognizing the need to take a strategic and systematic approach to risk management. To do that, they’re turning to risk modeling to help them navigate new forms of volatility and increase their access to capital, recognizing that the traditional tools and techniques are no longer sufficient.
Organizations with larger, more sophisticated risk management functions have adopted risk models to help them maximize the efficiency of their insurance programs — and save money in a hardening cycle that began in late 2019. Through the first quarter of 2021, global commercial insurance prices continued to climb, although there are some signs that rates may be moderating.
Faced with a firming market and an increase in loss frequency and severity, risk modeling provides a quantitative framework that leverages data and analytics to empower risk managers and other business leaders to make more informed decisions about their risk allocation and to better position them to take control of their risk.
And there’s plenty of room for more organizations to implement this tool. According to Aon’s 2019 Global Risk Management Survey, only one in five organizations said they used risk modeling. Part of the reason for this may be that the benefits of risk modeling are not widely understood. We break down common misperceptions and the benefits and limitations of risk modeling.
Myth No. 1: Risk Modeling Is Just an Exercise
The reality is that risk modeling can drive down an organization’s total cost of risk. That’s because risk models distill large amounts of internal and external data into a simple decision-making framework to maximize the efficiency of an insurance program. While traditional actuarial forecasting offers a point-estimate view of losses, risk modeling takes it to the next level by measuring the full range of losses associated with a company’s risk profile — allowing companies to plan ahead for myriad claims situations instead of reacting only after large losses occur. Rather than optimizing for individual coverage lines, a comprehensive risk model can provide a holistic view of the risk portfolio. What’s more, a risk model can overlay multiple insurance solutions to consider different coverage levels, limits, and retentions — enabling companies to compare alternative risk-transfer solutions and find innovative solutions for taking control of their risk.
Aon’s Risk Financing Decision Platform, for example, offers an intuitive framework for comparing insurance options that are consistent across coverage lines. The risk model evaluates each option according to the same two metrics: total cost of risk (premiums plus estimates of retained losses) and catastrophic total cost of risk. In this way, the model can objectively determine whether the cost savings of retaining more risk outweighs the additional risk exposure, based on the company’s risk tolerance.
Risk Modeling at Work
With the help of its broker, a multinational heavy-equipment manufacturer recently underwent risk modeling to optimize its balance between risk transfer and retention. The risk model established a comprehensive baseline view of the manufacturer’s current risk profile across all domestic and international coverage lines; considered its risk appetite (drawn from qualitative interviews with business unit managers up to the C-suite); forecasted the range of losses; and presented clear and accurate metrics to evaluate each program’s total cost of risk.
Armed with this information, risk leaders quickly realized they could afford to retain more risk and pull back from transferring risks to traditional markets. They set a threshold of $50 million for each coverage line based on their risk tolerance. Any amount above that would be transferred to the market and put into their captive, and any amount below would be retained.
The risk model provided a clear structure for risk leaders to align their insurance decisions to the enterprise risk appetite. The framework also made it easy for risk managers to articulate the decision and rationale to their stakeholders. The manufacturer is currently moving forward with implementation, which is expected to give it more control over insurance pricing and terms without adding significant catastrophic exposure.
Myth No. 2: All Data Is Equal
A risk model is only as good as its underlying data. The cleaner and better structured the data, the more accurate the risk-modeling outcomes. One of the biggest roadblocks to risk modeling is the absence of clean data — including loss and claims history and information on any other factors related to the risk profile. In addition to internal data, a risk model needs good data to forecast the external environment and insurance market. Solutions such as Aon’s Risk Financing Decision Platform deploy proprietary models that leverage industry information, along with brokerage insights on insurance market conditions, to ensure accurate and up-to-date risk forecasts.
Another crucial element of the data piece is engaging experienced risk modelers, who are increasingly difficult to find these days. Risk consulting firms that offer tailored solutions are led by teams with deep analytics proficiencies and years of industry experience. These professionals can help companies build a fit-for-purpose risk model and develop the right capabilities to unlock the power of data and analytics in risk management.
The Ins and Outs of Risk Model Data
- Historical loss data
- Risk exposures
- Risk probabilities
- Insurance coverage and premiums
- External environment conditions
- Total cost of risk = estimate of retained losses + premiums
- Catastrophic total cost of risk = potential cost of catastrophic risk
Myth No. 3: Risk Models Provide Perfect Insight
Risk models, like all statistical models, are imperfect representations of a complex and ever-changing world and should still be combined with other tools, including human judgment, for helpful insight. While we now have access to more data and analytical capabilities than ever, no algorithm can replace human thinking. It is critically important to compare the outputs for alignment with the reality of the loss environment and coverage terms. If the model produces $1 million claims annually but the history has claims of this size only every five years, the assumptions may need to be adjusted.
Qualitative inputs — such as a company’s tolerance and culture around risk — are inherently subjective and cannot be reduced to a set of numbers. A risk model serves as one input in the decision-making process and should be combined with other tools such as benchmarking, market research and human judgment.
The success of a risk-modeling program also depends on the “soft stuff” that surrounds the infrastructure. This includes the involvement of stakeholders who recognize the value of risk management (as opposed to risk avoidance) and managing the total cost of risk. An organizational backbone can encourage risk ownership and accountability, elevate risk managers to the leadership table, and embrace new risk techniques and solutions — such as risk modeling — to help better prepare for the risks of today and tomorrow.